Mapping the genomic mosaic of two ‘Afro-Bolivians’ from the ...be mainly of Native American...

RESEARCH ARTICLE Open Access

Mapping the genomic mosaic of two ‘Afro-Bolivians’from the isolated Yungas valleysJacobo Pardo-Seco1,2, Tanja Heinz1, Patricia Taboada-Echalar1, Federico Martinón-Torres2,3 and Antonio Salas1,2*

Abstract

Background: Unraveling the ancestry of ‘Afro-American’ communities is hampered by the complex demographicprocesses that took place during the Transatlantic Slave Trade (TAST) and the (post-)colonization periods. ‘Afro-Bolivians’ from the subtropical Yungas valleys constitute small and isolated communities that live surrounded bythe predominant Native American community of Bolivia. By genotyping >580,000 SNPs in two ‘Afro-Bolivians’, andcomparing these genomic profiles with data compiled from more than 57 African groups and other referenceancestral populations (n = 1,161 in total), we aimed to disentangle the complex admixture processes undergone by‘Afro-Bolivians’.

Results: The data indicate that these two genomes constitute a complex mosaic of ancestries that is approximately80 % of recent African origin; the remaining ~20 % being European and Native American. West-Central Africacontributed most of the African ancestry to ‘Afro-Bolivians’, and this component is related to populations living alongthe Atlantic coast (i.e. Senegal, Ghana, Nigeria). Using tract length distribution of genomic segments attributable todistinct ancestries, we could date the time of admixture in about 400 years ago. This time coincides with the maximumimportation of slaves to Bolivia to compensate the diminishing indigenous labor force needed for the development ofthe National Mint of Potosí.

Conclusions: Overall, the data indicate that the genome of ‘Afro-Bolivians’ was shaped by a complex process ofadmixture occurring in America among individuals originating in different West-Central African populations; theirgenomic mosaics received additional contributions of Europeans and local Native Americans (e.g. Aymaras).

Keywords: Transatlantic slave trade, Afro-Bolivians, Ancestry, Genome, SNPs

BackgroundDuring the Transatlantic Slave Trade (TAST) from the fif-teenth to the end of the nineteenth century, more than 12.5million enslaved Africans boarded ships that were destinedto the Americas. Most African slaves brought to the Ameri-cas came from West-Central Africa (~5.6 million) [1].African slaves were not distributed uniformly in theAmerican regions of interest. In Brazil, for example,most Africans were forced to migrate to the North-eastern part of the country, that is, to Pernambuco

and Bahia, because this region developed to becomeone of the most important sugar production zonesduring the TAST period [2]. Similarly, in Colombiathere was high demand for experienced gold minersespecially in the department of Chocó [3].During the last decade, geneticists have aimed to un-

ravel the complex patterns of admixture occurring inAmerica as a consequence of the European colonizationand the TAST. In particular, the analysis of mitochon-drial DNA (mtDNA) has been widely used in the lit-erature [4–10]. For instance, it is now known thatAfrican mtDNA lineages in America prevailed in theNortheast of Brazil [11], and are also widely distributed in‘Afro-Colombians’ of Chocó [12–14]. According to auto-somal DNA, many populations in Brazil are principally ofEuropean ancestry, but there exists a North–South gradi-ent towards an increased African ancestry in the North,Northeast, and Centre-West [15]. Similarly, in Colombia,

* Correspondence: [email protected] de Xenética, Departamento de Anatomía Patolóxica e CienciasForenses, and Instituto de Ciencias Forenses, Grupo de Medicina Xenómica(GMX), Facultade de Medicina, Universidade de Santiago de Compostela,Calle San Francisco s/n, C.P. 15872, Galicia, Spain2Grupo de Investigación en Genética, Vacunas, Infecciones y Pediatría(GENVIP), Hospital Clínico Universitario and Universidade de Santiago deCompostela (USC), Galicia, SpainFull list of author information is available at the end of the article

© 2016 Pardo-Seco et al. Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, andreproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link tothe Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver(http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.

Pardo-Seco et al. BMC Genomics (2016) 17:207 DOI 10.1186/s12864-016-2520-x

http://crossmark.crossref.org/dialog/?doi=10.1186/s12864-016-2520-x&domain=pdfmailto:[email protected]://creativecommons.org/licenses/by/4.0/http://creativecommons.org/publicdomain/zero/1.0/

a pronounced autosomal African admixture was observedin individuals from Chocó [16]. Moreno-Estrada et al. [17]investigated the population genetic history of theCaribbean by characterizing patterns of genome-widevariation. These authors found that admixed genomescan be traced back to distinct sub-continental sourcepopulations, even in situations where limited pre-Columbian Caribbean haplotypes survived.Very little attention has so far been given to historically

isolated populations whose history can also be traced backto the TAST, as is the case of the ‘Afro-Bolivian’ communityof Tocaña, located in the Nor (North) and the Sud (South)Yungas valleys of Bolivia. The Yungas are located in the De-partment of La Paz, in the passageway that connects theBolivian highlands and the tropics. During the colonialperiod, Spaniards initially used indigenous people as aworkforce to exploit the region’s mineral wealth at themines and Mint of Potosí, but soon they began to enslaveAfricans. Mortality rate of slaves was high in the regionmainly due to the fact that they were forced to work in a re-gion over 4,000 m above sea level and in very hard condi-tions [18]. The first African slaves in Potosí were recordedbeginning in 1549 [19]. Upper Peru, nowadays Bolivia, re-ceived a considerably number of African enslavedpeople at the beginning of the TAST from Senegam-bia; however, West-Central Africa became increasinglyimportant at the end of the sixteenth century. Rodri-guez [18] also mentions that at least 1,536 peoplewere brought to Bolivia from the Congo-Angola areaand Mozambique. The region around present-dayAngola might also have been an important source ofenslaved people for Spanish traders, because when theSpanish and Portuguese crown merged between 1580 and1640, many enslaved Africans from the hinterlands ofLuanda in Angola, which was kept by Portugal, wereforced to migrate to Spanish America [2, 20].African enslaved people from West-Central Africa prin-

cipally arrived in Upper Peru via Río de la Plata, and atleast between 1,500 and 3,000 Angolan enslaved peoplearrived at Río de la Plata every year during the first half ofthe seventeenth century [20]. It is also documented that in1807, there were 458 Africans in Potosí to work in themint [21]. Later, Spanish colonists began to use slaves inagricultural work in the tropical valleys. Rodriguez [18] re-ported that, in 1883, the enslaved ‘black’ population in theregion consisted of more than 6,000 people.Although ‘Afro-Bolivians’ were not included in the offi-

cial National Census of 1996, it was estimated at thattime that about 10,000 ‘Afro-Bolivians’ were mostly con-centrated in the Yungas provinces, mainly in rural townsand villages such as Coroico, Irupana, Tocaña, etc. Thissmall community adopted much of the technologicaland economic organization and cultural norms of thelocal indigenous Aymara [18]. Only very recently, the

reclaiming of a ‘Afro-Bolivian’ culture has begun toemerge in the region by the creation of cultural organi-zations aimed to recover their lost cultural identities.Some inferences on their origins have been made

based on linguistics [22]. There are unique words thatprobably derived from Kikongo, a language spoken inCongo. Furthermore, there exist two common Africansurnames in ‘Afro-Bolivian’ communities, Angola andMaconde, the latter of which might also be of Congoleseorigin [22].‘Afro-Bolivian’ groups have hitherto received very little

attention within the scientific community that investi-gates the history of the TAST; probably because theyconstitute a relatively small community surrounded by anumerous Aymara-speaking population, and becausethey live in a geographically remote region. Similarly,genetics research on Bolivian populations has primarilyfocused on the Native American indigenous popula-tion excluding people of African ancestry. So far,mtDNA lineages in the departments of Beni, Chuquisaca,Cochabamba, La Paz, Pando, and Santa Cruz, which aredistributed across three eco-geographically distinct re-gions (Andean, Sub-Andean, Llanos), have been shown tobe mainly of Native American ancestry [23–28]. Incontrast, the Y-chromosome shows an important con-tribution of European colonizers [28–31]. Likewise,autosomal DNA analysis (mainly carried out usingsmall panels of autosomal SNPs) shows a main NativeAmerican ancestry, although with an increased Europeanintrogression [23, 24]. In contrast, African ancestry wasobserved to be marginal in both mtDNA and autosomalDNA, respectively [23, 24, 32]. Only the community ofTocaña (Nor Yungas) still preserves the African geneticlegacy of the TAST [33] as inferred from the uniparentalmarkers and a few ancestry informative (autosomal)markers (AIMs).The aim of the present study is to provide a first

insight into the complex admixture processes experi-enced by individuals belonging to the geographicallyisolated ‘Afro-Bolivian’ community. In contrast to other‘Afro-American’ communities that admixed in complexdemographic circumstances (e.g. in the USA, Caribbean,Colombia, Brazil, etc.), ‘Afro-Bolivians’ from the Yungasvalleys constitute very isolated communities since theirinitial formation, and have remained surrounded bypeoples of main Native American ancestry. These‘Afro-Bolivians’ therefore constitute a sort of ‘geneticlaboratory’ to gain new insight into the TAST.

ResultsAnalysis of identity-by-state and multidimensional scalinganalysisThe genetic proximity of the two Tocaña profiles to dif-ferent sub-Saharan African groups can be studied by

Pardo-Seco et al. BMC Genomics (2016) 17:207 Page 2 of 12

examining the genetic distances in terms of Identity-by-State (IBS) values between these two Bolivians and eachof the population sample sets in Africa. These analyseswere carried out using separately the population setsfrom The 1000 Genomes Project (hereafter 1000G)and a large dataset of African populations (includingthose that contributed more slaves to the TAST). Theexploratory analysis carried out with 1000G samples(involving >500 K SNPs; see Material and Methodsfor details and Additional file 1 for the full list ofpopulation samples used for comparison) demon-strates that the two Bolivians have the highest IBSvalues with Africans (represented here by West-Central and East Africa), and the lowest values withnon-Africans (Additional file 2A). The second roundof analysis (Additional file 2B; 25 K SNPs), using apanel of 57 African datasets, indicates that the highestvalues of IBS for the two Tocaña are with the Yoruba(Nigeria); followed by a set of populations that are mainlyfrom West-Central Africa. The lowest IBS values arebetween the two Bolivians and North Africans.In order to better visualize the population relationships

between the two ‘Afro-Bolivians’ and the main continentalgroups, a MDS was carried out with the main continentalgroups represented in 1000G. Dimension 1 and Dimen-sion 2 (Additional file 3; >500 K SNPs) clearly separatesub-Saharans, East Asians, and Europeans in three tightclusters, each forming the vertex of an equilateral triangle.The remaining individuals (many of them admixed) plotin intermediate positions between the European edge andthe other two clusters, in good agreement with their docu-mented admixed ancestry; e.g. Puerto Rico shows a mainprojection from the European pole towards the Africanone, suggesting African admixture. In this scenario of con-tinental ancestries, the two ‘Afro-Bolivians’ clearly show aclose proximity to the sub-Saharans, in both the Dimen-sion 1 and the Dimension 2, with a minor projection to-wards the other two vertexes of the triangle.

Once the predominantly African nature of the two‘Afro-Bolivians’ was revealed, a second round of MDSwas carried out in order to investigate the relationshipsof the two Bolivians with different African sub-continental regions. The plot of Dimension 1 and Di-mension 2 shown in Additional file 4 (25 K SNPs) againdisplays a triangular arrangement, with the vertexesattracting South Africans (although very dispersed in thetriangle denoting admixture with other African groups),North Africans (with clear affinities to Europeans; hererepresented by CEU), and West-Central Africans. EastAfricans scatter along the side of the triangle thatconnects West-Central Africans to North Africans.South-East Africans plot very close to the vertex occu-pied by the West-Central African samples with someprojection into the South and East African poles as cor-responds with their admixed history [10]. The two indi-viduals from Tocaña locate in between the West-Centralpole and to East Africans (Additional file 4).Finally, PCAmask was run on a reduced set of popula-

tion samples. The first two dimensions (Fig. 1) indicatethat the African ancestry of the Tocaña haplotypes isclosely related to the main African group formed bythe Yoruba and the Luhya. The Native Americancomponent of the two Tocaña falls very close to themain Native American pole that includes the Aymaraand the Quechua.

Disentangling the African component of ‘Afro-Bolivians’PCAdmix was performed in order to reveal the genomearchitecture of the two Afro-Bolivians analyzed (Fig. 2).The analysis (>99 K SNPs) confirms the predominantlyAfrican component suggested by the MDS and indicatesa disperse genomic pattern for the European and NativeAmerican components, in good agreement with thehistorical process of admixture and incompatible withmodern events of admixture. Percentages of admix-ture in the two Tocaña calculated using PCAdmix

Fig. 1 PCAmask based on 99 K SNPs focused on the Native American (a) and the African ancestry (b) of the two Tocaña


were: (i) African ancestry: 78.1 for #TOC1 and 82.1 %for #TOC2; (ii) European ancestry: 14.8 for #TOC1and 2.6 % for #TOC2; and (iii) Native American an-cestry: 7.1 for #TOC1 and 15.3 % for #TOC2.An initial analysis performed with ADMIXTURE was

carried out between the two Yungas individuals and thepopulation sets from 1000G (>500 K SNPs). For anoptimum K = 5 (see Additional file 5 for other K values),this analysis (Fig. 3a) confirms that the two Bolivianshave a main African ancestry (#TOC1: 79.1 %; #TOC2:81.4 %; Additional file 6). Furthermore, according to thisanalysis, both Bolivians have more Native American(#TOC1: 11.6 %; #TOC2: 15.7 %) than European(#TOC1: 9.3 %; #TOC2: 2.9 %).Therefore, even though different reference population

datasets and algorithms were used, percentages of ad-mixtures obtained using PCAdmix and ADMIXTUREare broadly comparable for the three ancestral compo-nents. Both Tocaña have a similar proportion of Africanancestry (79–81 %); the main difference between them isthat #TOC2 has a larger proportion of Native Americanancestry than #TOC1, which is balanced by a lowerEuropean ancestry in #TOC2.Analysis of admixture was also carried out with the

large set of African populations (>25 K SNPs) andthe optimum K = 12 (Fig. 3b; Additional file 7; seeAdditional file 8 for other K values). Some findings ofthis exploratory analysis are suggestive. First, it indi-cates that the two Bolivians have one predominantcomponent (42–49 %) that coincides with the pre-dominant component in West-Central Africa (56 onaverage); in particular, in the Yoruba (Nigeria; 77),Hausa (Cameroon; 71), Brong (Ghana; 79), and Man-denka (Senegal; 77 %); Additional file 7. The second mostimportant component in the two Tocaña (23–25 %) isshared with a subset of East Africans; out of the wholeAfrican dataset this component reaches the maximumvalues in the Luhya (46–55 %; Additional file 7).

Disentangling the Native-American component of‘Afro-Bolivians’Additional analyses were carried out in order to fur-ther investigate the Native American component of‘Afro-Bolivians’ as suggested by ADMIXTURE andPCAdmix. This analysis is somehow blurred by: (i)the fact that all the Native American groups used asreference populations are also admixed to differentextents with Europeans and even with people of re-cent African ancestry [34], and (ii) the complex ad-mixture nature of the two Tocaña individualsinvestigated. The analysis was carried out using theunmasked (using the whole reference datasets; >99 KSNPs) and the masked (filtering out the non-NativeAmerican component of the individuals in the refer-ence populations; >92 K SNPs) data of the NativeAmerican groups analyzed by Reich et al. [34]; Fig. 4.ADMIXTURE reveals the presence of a unique Native

American component in the Tocaña, and a pattern thatis also present in the sub-Andean/Andean Bolivian pop-ulations (represented e.g. by Aymaras). Also remarkableis the fact that the Native American component of theTocaña does not share ancestry with the componentthat is more typical of the Llanos in Bolivia, representedhere by the neighboring population of the Surui fromthe Mato Grosso do Sul, Brazil. The genomic patternsdescribed are similar whether the unmasked (optimumK = 8; Fig. 4a) or masked Native American datasets(optimum K = 7; Fig. 4b) are used.

D-statisticsWe computed D-statistics using different combinationsof population datasets in order to formally test for theexistence of different admixture components in the twoTocaña as suggested from admixture analysis. First, wewere interested in testing if the Native American ances-try observed in the Tocaña was statistically significant.We therefore ran D(YRI, TOCi; XNA; Outgroup), with

Fig. 2 PCAdmix indicating the genomic architecture of the Tocaña individuals from the point of view of their main continental ancestries


Fig. 3 (See legend on next page.)


TOCi being each of the two Tocaña individuals, andXNA the different Native American datasets used in thepresent study. D-statistics were highly significant (>98 KSNPs) for all the Native American datasets (Fig. 5a) andthe two Tocaña. Interestingly, the highest significantvalue corresponded to the Aymara, followed by theQuechua (both from Bolivia).We next computed D(YRI, TOCi; XEU; Outgroup),

which used four different datasets as surrogate popula-tions to test for European admixture in the two Tocañas(XEU). D-statistics were again highly significant (>573 KSNPs) for the four different datasets used; Fig. 5b.Finally, we aimed to test for the presence of the East

Africa component in the Tocaña as suggested by theADMIXTURE analysis. We therefore computed D(YRI,TOCi; XEAF; Outgroup) and D(XEAF, TOCi; YRI; Out-group), with XEAF being the different East African data-sets available. As shown in Fig. 5c, D-statistic wasstatistically significant (>148 K SNPs) for most of thepopulation datasets from East Africa in both Bolivians(note that the threshold of significance was set to D-statistics < −2). This analysis was also very informativein showing that the Yoruba has a more importantcontribution to the Tocaña individuals than the EastAfrican populations, given that D-statistics showedmore negative values when using East Africans as ref-erence datasets; that is D(XEAF, TOCi; YRI; Outgroup).Also in good agreement with admixture proportions,D-statistics showed more negative values with #TOC2than with #TOC1 in D(XEAF, TOCi; YRI; Outgroup)indicating a larger West African ancestry of #TOC2.Finally, ADMIXTURE analysis indicated that, out ofthe East African population samples, the Bantu ethnicgroup of Luhya from Kenya seemed to show somegenetic affinities with the African component of thetwo Tocaña (Additional file 7); however, D-statisticswere not significant for this dataset.The fact that the two ‘Afro-Bolivians’ have a three-way

continental admixture might distort the results of the D-statistics. In particular, the seeming contribution in the twoTocaña from East Africa could be due to an artifact pro-voked by the presence of the European ancestry in both,East Africa and in the two Tocaña. In fact, admixture ana-lysis (Fig. 3) suggests the existence of such European com-ponent in both sets of populations. Therefore, inorder to test further for this hypothesis, we repeatedthe D-statistic analyses by first eliminating the non-

African component of the two Tocaña. For this ana-lysis, we used the output from PCAdmix after filter-ing the non-African chromosomal segments. D-statistics carried out on these masked Afro-Boliviangenomes finally revealed a lack of statistically signifi-cant contribution of East Africa to the two Tocaña ge-nomes (Fig. 5d).

Dating the time of admixtureThe output of PCAdmix (four haplotypes) was used todate the time of admixture as recorded in the two ‘Afro-Bolivian’ genomes and using the track length distributionof genomic segments inherited from different ancestralgroups [17, 35]. The data indicate that the most likelyscenario (−ln likelihood = 111.04) corresponds with amain admixture event occurring between a main Africanpopulation (88.1 %) with a minor European population(11.9 %) about 13 generations ago; which correspondwith 390 years ago (assuming 30 years per generation);Fig. 6. Right after (or even simultaneously), anotheradmixture event would occur with the local NativeAmerican population, then configuring the main genomearchitecture of present-day ‘Afro-Bolivian’ genomes.

Discussion and conclusionsThe results of the present study indicate that the two‘Afro-Bolivian’ genomes are a mosaic of different ances-tries. These genomes have a main sub-Saharan compo-nent (~80 %), with a minor contribution (~20 %)coming from Europe and Native America.The analyses suggest that West-Central Africa is the

region that contributed the main proportion of Africanancestry to ‘Afro-Bolivians’. The main signal comes fromthe Yoruba (Nigeria), but it seems also clear that thereare other populations in West-Central Africa, mainlyrepresenting the Atlantic coast (e.g. Ghana, Senegal,Nigeria), that could also have contributed to the genomeof the ‘Afro-Bolivians’ analyzed.Stimulated by the recent findings of Heinz et al. [33]

indicating the presence of mtDNA East African ances-try in the Yungas region, we tested for this ancestry inthe autosomal genome of the two ‘Afro-Bolivians’ ana-lyzed in the present study. In this regard, exploratoryADMIXTURE analysis suggested that the Luhya could,at least in part, explain the East African componentobserved in the Tocaña. Initial assessments derivedfrom the analysis of D-statistics also indicated a

(See figure on previous page.)Fig. 3 Bar-plot of individual ancestries as computed using the unsupervised clustering algorithm implemented in ADMIXTURE. a considers thepopulations in 1000G (K = 4 was the lowest cross validation value). b considers a wide set of African populations that represent mainsub-continental regions; one European [CEU] and one East Asian [CHB] sample were included for reference (K = 12 was the lowest crossvalidation value)


Fig. 4 Analysis of admixture of the two Tocaña individuals with unmasked (a) and masked (b) Native American datasets


seeming contribution of East African ancestry to the‘Afro-Bolivian’ genomes. However, further analysis demon-strated that this signal could also be explained by compu-tational artifacts provoked by the presence of Europeanancestry in the genome of the two Tocaña and in the EastAfrican populations.While the African ancestry of ‘Afro-Bolivians’ consti-

tutes a complex mosaic of African ancestries, their NativeAmerican component is probably more homogeneous,

even though the analysis is very complicated by theadmixed nature of all reference Native American popula-tions [34]. The Native American component in theTocaña genomes is compatible with a major contributionof the surrounding Aymara population (with which it issupposed to be admixed), and incompatible with admix-ture with populations that are more representative of theLlanos from Bolivia (characterized here by the neighbor-ing Surui population).

Fig. 5 D-statistics computed on different population contexts and considering the two Tocaña individuals separately and Native Americansamples (a) European samples (b), East African samples (c), and East African samples but eliminating from the Tocaña genomes the non-Africanancestry (d). Note that the values are not comparable between the three different figures; apart from using different scales the three analyses arebased on different amounts of SNPs (see text for more information)


The time of admixture, as reflected in the two ‘Afro-Bolivian’ genomes, could be derived from the tractlength distribution of genomic segments inherited fromdistinct ancestries. The data indicate a main admixtureevent occurring at about 400 years ago between Africans,Europeans and Native Americans (Fig. 6). Interestingly,this event coincided with the period of maximum import-ation of slaves to Bolivia, which played a crucial role bycompensating the diminishing indigenous labor forceneeded for the development of the National Mint ofPotosí. The Mint of Potosí was the origin of most of thesilver shipped to Spain at the time; the initial Mint datedto 1572, and the second factory was set in 1757 and wasused till 1773.Among the limitations of the present study is the fact

that the reference populations in Africa, America andEurope, used in the present analyses may not representthe patterns of variability existing at the time of theTAST. In addition, although we compiled informationon hundreds of individuals representing dozens ofNative American and European populations, and morespecially population samples in Africa, the size of thereference populations might still be small. In the future,an improved database could help corroborate the presentfindings and add more details on the demographic pro-cesses that led to the configuration of present-day ‘Afro-Bolivian’ genomic architecture (and ‘Afro-Americans’ ingeneral).With the arrival of new genotyping and computational

technologies it is now possible to map with muchbetter resolution the complex genomic mosaic of‘Afro-Americans’. The case of ‘Afro-Bolivians’ is par-ticularly interesting owing to their strong isolationfrom other African and Native American communi-ties. Disentangling history of ‘Afro-Americans’ is ham-pered by the fact that many enslaved African peopletook unusual journeys after their enslavement. For ex-ample, Hans Jonathan, born as a slave on the DanishCaribbean island of St. Croix, was forced later in his life to

migrate with his master to Copenhagen, Denmark, butwhen slavery was abolished in Denmark he was sentencedto go back as a slave to St. Croix. Jonathan escaped fromslavery and finally landed in Iceland [36].The present article shows that genomics is useful to

reconstruct the individual ancestries of ‘Afro’ commu-nities living in the Americas, accounting for the com-plex processes of admixture occurring since theirarrival to the American continent. Most of the previ-ous attempts were based on the analysis of uniparen-tal markers [4, 6, 7, 9, 13, 14, 37, 38]. Although thesemarkers are phylogeographically very informative at apopulation scale, they have limitations when dealingwith individual ancestries (as it is the case of the twoTocaña individuals analyzed here), given that theirmtDNA and Y-chromosome trace back to only twoancestors among the many thousands that could the-oretically have contributed to their full genomes sincethe arrival of Europeans and the TAST into Bolivianterritories.

MethodsSamplingA collection of 105 saliva samples recruited from theYunga’s region of Bolivia was previously characterizedfor mtDNA and a panel of Ancestry InformativeMarkers or AIMs [33]. A small subset of these sam-ples was additionally analyzed for a selection of Y-chromosome markers [29]. For genome-wide analysis,we initially selected a subset of these samples (n = 9)that showed a large membership of African ancestryin the analysis of a small panel of AIMs. We obtainedwritten informed consent for all the donors prior theresearch, which includes consent for publication indi-vidual data. Rights of participants were safeguardedduring the research and their identity was protected.The study was approved by the ethical commission ofthe Universidade de Santiago de Compostela (Galicia;Spain) which depends on Ethics Committee of Galicia(Spain) and by the Ethics Committee of UniversidadMayor, Real y Pontificia de San Francisco Xabier deChuquisaca (Bolivia). The study conforms with all ap-plicable Spanish normative, this is to say, Law forBiomedical Research (Law 14/2007–3 of July), Law41/2002 of Autonomy of the Patient, Decree SAS/3470/2009 for observational studies and Law 15/1999of Data Protection.

GenotypingSamples were genotyped using the Axiom® Genome-Wide Human Origins 1 Arrays at the Centro Nacionalde Genotipado (CEGEN) of Santiago de Compostela(Spain). After processing the samples, the AffymetrixGenotyping Console™ (GTC) 4.1.2 Software package was

Fig. 6 Time and model of admixture of ‘Afro-Bolivians’. The area ofthe pie charts above the migration model are proportional to theestimated number of migrants being introduced at each point intime (indicated by black arrows), as done in [35]


used to generate QC metrics and genotype calls follow-ing commercial recommendations. This procedure in-cludes generation of sample dish QC (DQC) values.Samples with a DQC value lower than the default DQCthreshold of 0.82 were dropped from the study. The“AxiomGT1_all” algorithm was used for genotyping thisgroup of high quality samples using all SNPs on thearray. Unfortunately, not all the samples initially selectedfor genotyping contained enough DNA (and/or of therequired quality) for genome-wide SNP genotyping. Fi-nally, only the analysis of two of the samples succeeded,with a total of 580,144 SNPs genotyped. These two sam-ples have been recruited in the locality of Tocaña, andare referred to here as #TOC1 and #TOC2 (originalsample ID: #Tocana301 and #T303; respectively).

Reference populationsWe intersected the SNP data obtained from the two‘Afro-Bolivians’ with genome population data down-loaded and processed from different SNP repositories;Additional file 1. Comparative population analyses werecarried out using different sets of data. The differentcombination of populations sets used for each analysis isalso indicated in Additional file 1.Data from The 1000 Genomes Project (hereafter

1000G) provides the dataset with the largest SNP over-lap with the Tocaña data. The data from 1000G wasretrieved from the original repository as done in Pardo-Seco et al. [39]; these analyses involved 580,144 SNPs(1.7 % of missing data).The lowest SNP overlap occurred when intersecting

the Tocaña SNP set with a large compilation of popula-tion samples that includes 1,161 Africans (representing57 different populations/ethnic groups); these analysesinvolved 25,192 SNPs (0.7 % of missing data).Some ad hoc analyses were carried out using selected

sub-sets of populations with the aim to increase thenumber of SNPs (specific numbers are given in the maintext where relevant). For instance, analyses of the NativeAmerican component observed in the two Tocaña indi-viduals were carried out using the unmasked (0.2 miss-ing data) and masked (7.8 % missing data) dataset fromReich et al. [34]; these datasets include 99,378 and92,619 SNPs, respectively.

Statistical analysisPLINK [40] was used to compute IBSe values from SNPdata. Before proceeding with all population analyses, wetested the two Tocaña genomic profiles for potentialclose relationships given that both individuals were sam-pled in the same location. We followed the proceduredetailed in Gómez-Carballa et al. [41]. Identity-by-descent (IBD) values for this pair of genetic profiles arenot compatible with close familial relationships after

based on statistical tests proposed before [42, 43]; P-value = 1.Admixture components of the two Bolivian individuals

were performed using the ADMIXTURE software [25],which uses a maximum likelihood estimation of individ-ual ancestries from multi-locus SNP data.Local ancestry assignment of ancestry-specific haplo-

types across the genome was carried out using PCAdmix1.0 [44]. PCAdmix requires haplotype data; therefore, inorder to prepare input files for PCAdmix, SNP unphaseddata were imputed and haplotypes were built usingBeagle 3.3.2. [45]. For this particular analysis, we se-lected populations from 1000G complemented withsome Native American groups analyzed in Reich et al.[34] in order to represent the three main ancestralgroups of the Tocaña: Yoruba and Luhya (representingAfrican ancestry), CEU (representing European ancestry),and Aymara, Guahibo, Guarani, Hullicje Maya, Mixe,Piapoco, Quechua, Ticuna, Toba, and Wichi (representingNative American ancestry). We assigned genomic seg-ments to the three possible ancestries following a poster-ior probability threshold of 0.8.In order to discriminate among clusters of genetic

variation in the population sets analyzed, multidimen-sional scaling (MDS) was carried out on a matrix of pair-wise individual IBS values. MDS was performed usingthe function cmdscale (library stats) from R (http://www.r-project.org). PCAmask [17] was additionallycarried out on genomic segments assigned to NativeAmerican, African, or European ancestry. This analysisuses a haplotype-based algorithm and therefore includessegments of the genome that are heterozygous for ances-try (thus minimizing missing data). Given that PCAmaskuses the output of PCAdmix, we run PCAmask usingthe same sub-set of samples.We additionally computed the four-population test,

implemented as D-statistics [46], in order to determinethe relationship between the two Tocaña SNP profilesindividually vs. different population sets [34, 47–49].This is a formal test for admixture that measure allelefrequency correlations among populations; it providesstatistical evidence of admixture and information on thedirectionality of the gene flow. D-statistic was computedusing the weighted block jackknife procedure (block sizeof 5 MB) [47].For the computation of D-statistics, we built an out-

group that is symmetrically related to all modern humanpopulation groups, by creating an individual profile pos-sessing the ancestral alleles at all sites. The use of thisartificial outgroup allowed us to infer D-statistics, thusensuring that there is no differential gene flow betweenthe outgroup and the population sets used; this is par-ticularly important in our population context given thatthe three main continental ancestries are present in our


http://www.r-project.orghttp://www.r-project.org

sample sets. The construction of this outgroup involvedthe following steps. First, an auxiliary ancestral sampleVCF file was generated containing all the ancestral al-leles as alternative alleles. Second, all the chromosomesites from 1000G VCF files where parsed looking for“AA” codes, the ancestral alleles where detected and thealternative alleles where substituted with the ancestralalleles. Only the first four VCF columns where kept,then the ancestral allele was added, and the rest of thecolumns where filled with single dots. The genotype col-umn was filled with reference homozygous (0/0) in caseswhen only the reference was reported, and with alterna-tive heterozygous (0/1) when an ancestral allele differentfrom the reference was reported.Finally, timing of admixture was estimated by way of

analyzing the tract length of chromosomal segmentsattributed to different ancestries, using the softwareTracts [35]; see also Moreno-Estrada et al. [17].

Additional files

Additional file 1: Population datasets used in the present study. (XLS 66 kb)

Additional file 2: Average IBS values between the two individuals fromTocaña and individuals from various continental regions represented in(A) 1000G, and (B) a large dataset of African regions. (TIFF 10002 kb)

Additional file 3: MDS of the two Tocaña individuals vs. the populationsets from 1000G representing the main continental groups. See Additionalfile 1 for more information on population datasets. (TIFF 2502 kb)

Additional file 4: MDS of Tocaña against 57 datasets representingdifferent sub-continental African regions. One population of Europeanancestry (CEU) from 1000G were used for reference. See Additional file 1for more information on population datasets. (TIFF 27503 kb)

Additional file 5: Analysis of admixture as in Fig. 3a for additional Kvalues. (TIFF 15628 kb)

Additional file 6: Ancestry memberships as computed using theunsupervised clustering algorithm from ADMIXTURE using the 1000Gdatasets and the Tocaña, for an optimum value of K = 5. (XLSX 57 kb)

Additional file 7: Ancestry memberships as computed using theunsupervised clustering algorithm from ADMIXTURE and a large datasetrepresenting main sub-continental African regions and the Tocaña, for anoptimum value of K = 12. (XLSX 14 kb)

Additional file 8: Analysis of admixture as in Fig. 3b for additional Kvalues. (TIFF 31254 kb)

Competing interestsThe authors declare that they have no competing interests.

Authors’ contributionsTH and AS conceived the study and FMT participated in its design. AS, PTEand FMT provided reagents, materials and analysis tools. JPS, TH, PTE, and ASanalyzed the data. AS drafted the article and all the authors critically revisedthe manuscript and gave final approval of the version to be published.

AcknowledgementsWe would like to thank J. Amigo for his assistency in preparing the outgroupfor the differente D-statistics analyses carried out in the present study. Theresearch leading to these results has received funding from the PeopleProgram (Marie Curie Actions) of the European Union’s Seventh FrameworkProgram FP7/2007–2013/under REA grant agreement no. 290344, from the“Ministerio de Ciencia e Innovación” (SAF2011–26983), the “Plan Galego IDT”(EM 2012/045) and the grant from the “Sistema Universitario Gallego-

Modalidad REDES (2012-PG226) from the Xunta de Galicia (A.S.). F.M-Treceived support from the grant “ISCIII/INT14/00245/ CofinanciadoFEDER".

Author details1Unidade de Xenética, Departamento de Anatomía Patolóxica e CienciasForenses, and Instituto de Ciencias Forenses, Grupo de Medicina Xenómica(GMX), Facultade de Medicina, Universidade de Santiago de Compostela,Calle San Francisco s/n, C.P. 15872, Galicia, Spain. 2Grupo de Investigación enGenética, Vacunas, Infecciones y Pediatría (GENVIP), Hospital ClínicoUniversitario and Universidade de Santiago de Compostela (USC), Galicia,Spain. 3Infectious Diseases and Vaccines Unit, Department of Pediatrics,Hospital Clínico Universitario de Santiago, Santiago de Compostela, Galicia,Spain.

Received: 6 August 2015 Accepted: 24 February 2016

References1. Eltis D. A brief overview of the Trans-Atlantic Slave Trade. Voyages: The

Trans-Atlantic Slave Trade Database: http://www.slavevoyages.org; 2008.2. Klein HS. The Atlantic Slave Trade (New approaches to the Americas).

Cambridge: Cambridge University Press; 2010.3. Davis DJ. Beyond Slavery: The multilayered legacy of Africans in Latin

America and the Caribbean. INC, USA: Rowman & Littlefield Publishers; 2007.4. Salas A, Richards M, Lareu MV, Scozzari R, Coppa A, Torroni A, et al. The

African diaspora: mitochondrial DNA and the Atlantic slave trade. Am JHum Genet. 2004;74(3):454–65.

5. Salas A, Torroni A, Richards M, Quintana-Murci L, Hill C, Macaulay V, et al.The phylogeography of mitochondrial DNA haplogroup L3g in Africa andthe Atlantic slave trade. Am J Hum Genet. 2004;75:524–6.

6. Salas A, Carracedo Á, Richards M, Macaulay V. Charting the Ancestry ofAfrican Americans. Am J Hum Genet. 2005;77(4):676–80.

7. Alves-Silva J, da Silva SM, Guimaraes PE, Ferreira AC, Bandelt H-J, Pena SD,et al. The ancestry of Brazilian mtDNA lineages. Am J Hum Genet.2000;67(2):444–61.

8. Ely B, Wilson JL, Jackson F, Jackson BA. African-American mitochondrialDNAs often match mtDNAs found in multiple African ethnic groups.BMC Biol. 2006;4:34.

9. Stefflova K, Dulik MC, Barnholtz-Sloan JS, Pai AA, Walker AH, Rebbeck TR.Dissecting the within-Africa ancestry of populations of African descent inthe Americas. PLoS One. 2011;6(1):e14495.

10. Salas A, Richards M, De la Fé T, Lareu MV, Sobrino B, Sánchez-Diz P, etal. The making of the African mtDNA landscape. Am J Hum Genet.2002;71(5):1082–111.

11. Parra FC, Amado RC, Lambertucci JR, Rocha J, Antunes CM, Pena SD.Color and genomic ancestry in Brazilians. Proc Natl Acad Sci U S A.2003;100(1):177–82.

12. Rojas W, Parra MV, Campo O, Caro MA, Lopera JG, Arias W, et al. Geneticmake up and structure of Colombian populations by means of uniparentaland biparental DNA markers. Am J Phys Anthropol. 2010;143(1):13–20.

13. Salas A, Richards M, Lareu MV, Sobrino B, Silva S, Matamoros M, et al.Shipwrecks and founder effects: divergent demographic histories reflectedin Caribbean mtDNA. Am J Phys Anthropol. 2005;128:855–60.

14. Salas A, Acosta A, Álvarez-Iglesias V, Cerezo M, Phillips C, Lareu MV, et al.The mtDNA ancestry of admixed Colombian populations. Am J Hum Biol.2008;20:584–91.

15. Lins TC, Vieira RG, Abreu BS, Grattapaglia D, Pereira RW. Genetic composition ofBrazilian population samples based on a set of twenty-eight ancestryinformative SNPs. Am J Hum Biol. 2010;22(2):187–92.

16. Ibarra A, Freire-Aradas A, Martínez M, Fondevila M, Burgos G, Camacho M, etal. Comparison of the genetic background of different Colombianpopulations using the SNPforID 52plex identification panel. Int J Legal Med.2014;128(1):19–25.

17. Moreno-Estrada A, Gravel S, Zakharia F, McCauley JL, Byrnes JK, Gignoux CR,et al. Reconstructing the population genetic history of the Caribbean.PLoS Genet. 2013;9(11):e1003925.

18. Rodríguez RJ. The Afro populations of America’s southern cone: organization,development, and culture in Argentina, Bolivia, Paraguay, and Uruguay. Oxford,England: Towman & Littlefield Publishers, Inc; 2001.


dx.doi.org/10.1186/s12864-016-2520-xdx.doi.org/10.1186/s12864-016-2520-xdx.doi.org/10.1186/s12864-016-2520-xdx.doi.org/10.1186/s12864-016-2520-xdx.doi.org/10.1186/s12864-016-2520-xdx.doi.org/10.1186/s12864-016-2520-xdx.doi.org/10.1186/s12864-016-2520-xdx.doi.org/10.1186/s12864-016-2520-xhttp://www.slavevoyages.org

19. Davies CEB. Encyclopedia of the African diaspora: origins, experiences, andculture. Santa Barbara, California: ABC-CLIO; 2008.

20. Hall EJ. Slavery and African Ethnicities in the Americas: Restoring the links.USA: The University of North Carolina Press; 2005.

21. Lipski JM. Afro-Bolivian Spanish and Helvécia Portugues: semie-Creoleparallels. Papia. 2006;16:96–116.

22. Lipski JM. Afro-Bolivian language today: the oldest surviving Afro-Hispanicspeech community. Afro-Hispanic Review. 2006;25(1):179.

23. Heinz T, Álvarez-Iglesias V, Pardo-Seco J, Taboada-Echalar P, Gómez-Carballa A,Torres-Balanza A, et al. Ancestry analysis reveals a predominant NativeAmerican component with moderate European admixture in Bolivians.Forensic Sci Int Genet. 2013;7(5):537–42.

24. Taboada-Echalar P, Álvarez-Iglesias V, Heinz T, Vidal-Bralo L, Gómez-Carballa A,Catelli L, et al. The genetic legacy of the pre-Colonial period in contemporaryBolivians. PLoS One. 2013;8(3):e58980.

25. Afonso-Costa H, Carvalho M, Lopes V, Balsa F, Bento AM, Serra A, et al.Mitochondrial DNA sequence analysis of a native Bolivian population.J Forensic Leg Med. 2010;17(5):247–53.

26. Bert F, Corella A, Gene M, Perez-Perez A, Turbon D. Mitochondrial DNAdiversity in the Llanos de Moxos: Moxo, Movima and Yuracare Amerindianpopulations from Bolivia lowlands. Ann Hum Biol. 2004;31(1):9–28.

27. Bert F, Corella A, Gene M, Pérez-Pérez A, Turbón D. Major mitochondrialDNA haplotype heterogeneity in highland and lowland Amerindianpopulations from Bolivia. Hum Biol. 2001;73(1):1–16.

28. Gayà-Vidal M, Moral P, Saenz-Ruales N, Gerbault P, Tonasso L, Villena M, etal. mtDNA and Y-chromosome diversity in Aymaras and Quechuas fromBolivia: Different stories and special genetic traits of the Andean Altiplanopopulations. Am J Phys Anthropol. 2011;145(2):215–30.

29. Cárdenas JM, Heinz T, Pardo-Seco J, Álvarez-Iglesias V, Taboada-Echalar P,Sánchez-Diz P, et al. The multiethnic ancestry of Bolivians as revealed by theanalysis of Y-chromosome markers. Forensic Sci Int Genet. 2014;14:210–8.

30. Tirado M, López-Parra AM, Baeza C, Bert F, Corella A, Pérez-Pérez A, et al.Y-chromosome haplotypes defined by 17 STRs included in AmpFlSTR YfilerPCR Amplification Kit in a multi ethnical population from El BeniDepartment (North Bolivia). Leg Med (Tokyo). 2009;11(2):101–3.

31. Vullo C, Gomes V, Romanini C, Oliveira AM, Rocabado O, Aquino J, et al.Association between Y haplogroups and autosomal AIMs reveals intra-population substructure in Bolivian populations. Int J Legal Med. 2014.

32. Watkins WS, Xing J, Huff C, Witherspoon DJ, Zhang Y, Perego UA, et al.Genetic analysis of ancestry, admixture and selection in Bolivian andTotonac populations of the New World. BMC Genet. 2012;13:39.

33. Heinz T, Cárdenas JM, Álvarez-Iglesias V, Pardo-Seco J, Gómez-Carballa J,Santos C, et al. The genomic legacy of the Transatlantic Slave Trade in theYungas valley of Bolivia. PLoS One. 2015;10(8):e0134129.

34. Reich D, Patterson N, Campbell D, Tandon A, Mazieres S, Ray N, et al.Reconstructing Native American population history. Nature.2012;488(7411):370–4.

35. Gravel S. Population genetics models of local ancestry. Genetics.2012;191(2):607–19.

36. Loftsdóttir K, Pálsson G. Black on White: Danish Colonialism, Iceland and theCaribbean. New York: Springer; 2013.

37. Rodas C, Gelvez N, Keyeux G. Mitochondrial DNA studies show asymmetricalAmerindian admixture in Afro-Colombian and Mestizo populations. Hum Biol.2003;75(1):13–30.

38. Brucato N, Cassar O, Tonasso L, Tortevoye P, Migot-Nabias F, Plancoulaine S,et al. The imprint of the Slave Trade in an African American population:mitochondrial DNA, Y chromosome and HTLV-1 analysis in the Noir Marronof French Guiana. BMC Evol Biol. 2010;10:314.

39. Pardo-Seco J, Gómez-Carballa A, Amigo J, Martinón-Torres F, Salas A. Agenome-wide study of modern-day Tuscans: revisiting Herodotus’s theoryon the origin of the Etruscans. PLoS One. 2014;9(9):e105920.

40. Purcell S, Neale B, Todd-Brown K, Thomas L, Ferreira MA, Bender D, et al.PLINK: a tool set for whole-genome association and population-basedlinkage analyses. Am J Hum Genet. 2007;81(3):559–75.

41. Gómez-Carballa A, Pardo-Seco J, Fachal L, Vega A, Cebey M, Martinón-Torres N,et al. Indian signatures in the westernmost edge of the European romanidiaspora: new insight from mitogenomes. PLoS One. 2013;8(10):e75397.

42. Stevens EL, Heckenberg G, Roberson ED, Baugher JD, Downey TJ, Pevsner J.Inference of relationships in population data using identity-by-descent andidentity-by-state. PLoS Genet. 2011;7(9):e1002287.

43. Lee WC. Testing the genetic relation between two individuals using a panelof frequency-unknown single nucleotide polymorphisms. Ann Hum Genet.2003;67(Pt 6):618–9.

44. Henn BM, Botigué LR, Gravel S, Wang W, Brisbin A, Byrnes JK, et al. Genomicancestry of North Africans supports back-to-Africa migrations. PLoS Genet.2012;8(1):e1002397.

45. Browning SR, Browning BL. Rapid and accurate haplotype phasing andmissing-data inference for whole-genome association studies by use oflocalized haplotype clustering. Am J Hum Genet. 2007;81(5):1084–97.

46. Green RE, Krause J, Briggs AW, Maricic T, Stenzel U, Kircher M, et al. A draftsequence of the Neandertal genome. Science. 2010;328(5979):710–22.

47. Durand EY, Patterson N, Reich D, Slatkin M. Testing for ancient admixturebetween closely related populations. Mol Biol Evol. 2011;28(8):2239–52.

48. Schroeder H, Avila-Arcos MC, Malaspinas AS, Poznik GD, Sandoval-Velasco M,Carpenter ML, et al. Genome-wide ancestry of 17th-century enslaved Africansfrom the Caribbean. Proc Natl Acad Sci U S A. 2015;112(12):3669–73.

49. Lazaridis I, Patterson N, Mittnik A, Renaud G, Mallick S, Kirsanow K, et al.Ancient human genomes suggest three ancestral populations for present-day Europeans. Nature. 2014;513(7518):409–13.

• We accept pre-submission inquiries • Our selector tool helps you to find the most relevant journal• We provide round the clock customer support • Convenient online submission• Thorough peer review• Inclusion in PubMed and all major indexing services • Maximum visibility for your research

Submit your manuscript atwww.biomedcentral.com/submit

Submit your next manuscript to BioMed Central and we will help you at every step:


AbstractBackgroundResultsConclusions

BackgroundResultsAnalysis of identity-by-state and multidimensional scaling analysisDisentangling the African component of ‘Afro-Bolivians’Disentangling the Native-American component of ‘Afro-Bolivians’D-statisticsDating the time of admixture

Discussion and conclusionsMethodsSamplingGenotypingReference populationsStatistical analysis

Additional filesCompeting interestsAuthors’ contributionsAcknowledgementsAuthor detailsReferences

Date post:	14-Feb-2021
Category:	Documents
Upload:	others
View:	0 times
Download:	0 times

Mapping the genomic mosaic of two ‘Afro-Bolivians’ from the ...be mainly of Native American...

Documents